Opinion
No. 7380.
Decided March 17, 1941.
Appeal from the District Court of the United States for the District of Columbia.
Action by Arnold Poulsen and another against Conway P. Coe, Commissioner of Patents, for an order authorizing defendant to issue a patent on plaintiffs' application for letters patent. From a decree dismissing the bill, plaintiffs appeal.
Decree affirmed in part and reversed in part, and case remanded for further proceedings.
Paul Kolisch, of New York City, and Ralph B. Stewart, of Washington, D.C., for appellants.
W.W. Cochran, Sol., United States Patent Office, of Washington, D.C., for appellee.
Before STEPHENS, EDGERTON, and RUTLEDGE, Associate Justices.
This is an appeal from a decree of the District Court of the United States for the District of Columbia dismissing, after a hearing on the merits, a bill in equity seeking, under Rev.Stat. § 4915, 35 U.S.C.A. § 63, an order authorizing the appellee Commissioner of Patents to issue to the appellants Poulsen and Petersen a patent on their application for letters patent, Serial No. 563,447, filed December 17, 1931, with claims 49, 70, 72, 73, 74, 76, 77, 78 and 79. Claims 72, 73 and 74 are process claims; the others are claims for a machine, informally denominated a system or apparatus. All of the claims relate to the recording of sound on films such as are used in the making of talking motion pictures. All, except claim 70, were rejected by the Commissioner as lacking invention over the prior art; claim 70, as not sufficiently definite to meet the requirements of Rev.Stat. § 4888, 35 U.S.C.A. § 33. The District Court in dismissing the appellants' bill apparently accepted the views of the Patent Office with respect to each of the claims. The references relied upon in the Patent Office were:
Sacia 1,623,756 April 5, 1927 Hanna 1,888,724 November 22, 1932 Robinson 1,854,159 April 12, 1932 Hewlett 1,853,812 April 12, 1932
Sacia was not urged at the trial below. On this appeal the reliance in the Commissioner's brief is upon Robinson and Hewlett alone. In this opinion we shall therefore disregard both Sacia and Hanna.
The basic process of sound recording in connection with the production of motion pictures is the transformation of tonal variations produced by a sound source into variations in the magnitude of current in an electric circuit, and then into light variations which are recorded by exposure to them of a moving logitudinal strip of film called a sound track. An understanding of this case requires a brief description of the art as it existed prior to the references relied on in the appeal, and as it existed in view of those references, as well as a description of the appellants' apparatus and method. Such description can best be made in terms of Illustrations I and II below. The art prior to Robinson and Hewlett: As sound to be recorded emanates from its source, it produces pressure vibrations usually spoken of as sound waves which are intercepted by a microphone. This, by means of a diaphragm, converts the mechanical energy of the sound waves into electric energy in the form of an alternating current with vibrational characteristics which are identical with those of the sound waves. The conduits of this alternating current appear in Illustration I at the points marked a a'.
Illustration I is based on Exhibit 1 in the record; Illustration II, on Exhibit 3. Illustration II includes only Figures A, C and D of Exhibit 3. Figure B is omitted because it is illustrative of Hanna.
These conduits, passing first through an amplifier, connect the microphone to an oscillograph loop b b' to which is attached a mirror c. The oscillation of this mirror, and, in consequence, of a light beam reflected from it in the manner described below, varies in accordance with the changing magnitude of the alternating current vibrations — which itself of course varies in accordance with the variation in amplitude of the sound waves which enter the microphone. The beam of light to be reflected by the mirror upon the moving film is produced by a lamp whose rays pass through a rectangular mask containing a condenser lens which focuses the bundle of rays upon the mirror. When reflected by this mirror toward the film the light bundle thus produced strikes a screen in which there is a horizontal slit, so that there flows on toward the film only that portion of the bundle of light which passes through the slit; and this portion of the bundle is itself, through an objective lens, focused into a fine, transversely disposed line of light which vibrates horizontally upon the sound track, d d', moving downward in a perpendicular plane. In this manner the recording end of the line of light produces upon the film when developed a negative representation of the light vibrations — which themselves correspond exactly to the original vibrations from the sound source. The squares e e' shown in Illustration I represent sprocket holes by means of which the film is fitted over moving sprockets which carry it downward. The shaded surfaces are those portions of the film upon which the visual images are recorded.
An oscillograph is an instrument for the delineation of alternating current wave forms. It consists of two metal conduits which pass side by side under tension between the poles of an electromagnet. The current is led down one strip and up the other so that under and in accordance with the fluctuations of the current the strips move in opposite directions at right angles to their length and thus cause to oscillate a small mirror bridged across the strips. This mirror reflects a beam of light horizontally onto a vertically disposed photographic film which moves downward and thus at right angles to the light beam and which through exposure to the light beam thus delineates variations in the current. 16 Encyclopaedia Britannica (14th ed., 1929, 1932) 949.
The method of sound recording here described is known in the art as variable area, as distinguished from variable density, sound recording. As explained by the examiner: "Two types of records may be made. A variable density record [is one] in which the light modulator causes a constant area projected on the film to vary in light intensity linearly with the sound intensity and frequency. A variable area record [is one] in which the light modulator causes a constant intensity light projected on the film to vary in area linearly with the sound intensity and frequency." See "Examiner's Statement on Appeal before the Board of Appeals," R. 33. See also Types of Film Recording, L.E. Clark and John K. Hilliard, Chapter III, p. 24, Motion Picture Sound Engineering (D. Van Nostrand Company, Inc., 1938).
The sound track is shown in the Illustration in the form of a positive rather than a negative. Actually in the art, as it is now being described, the line of light produces a film which when developed is a negative — so that that portion of the film which is exposed to the variations of the line of light is opaque and the remainder transparent.
In the art prior to Hewlett and Robinson, the oscillograph mirror was so adjusted mechanically that when no sound was being produced and therefore no current was passing through the oscillograph loop, the recording end of the line of light was placed exactly in the middle of the sound track. In Figure A of Illustration II there appears graphically the vibration pattern produced by the apparatus thus far described, with the oscillograph mirror set as just stated.
It will be noted that the peaks of the waves described by the recording end of the line of light are symmetrical with respect to a median line or axis a a' — sometimes referred to as the "zero line," which is midway between the sides of the sound track. It is thus shown that under conditions of silence one half the sound track is exposed and the other half unexposed. That is to say, in this stage of the art, when there was no oscillation of the oscillograph mirror, the transversely disposed line of light
upon the sound track would expose only that portion of the latter designated b b' in Figure A.
In explaining this, the appellant Poulsen testified: "It will be seen that we start in the middle of the sound track. A line of light was in those days always so adjusted that the end of the light line was in the middle of the sound track, so that when there would be no sound half of the track would be exposed and the other half would be unexposed, and therefore after development half of the track would be black and the other half would be transparent."
But in this stage of the art there was a serious problem. When sound recorded upon a sound track in light variations as above described is to be reproduced for the benefit of a theatre audience, a fixed light source is sharply focused upon and shines through a positive, produced by printing the developed film record negative, while the former is moved in front of a photoelectric cell. The cell thus "sees" the linear variations in light and functions to transmute them into corresponding electric variations which are amplified to a desired degree and are then led into and actuate the diaphragm of the conventional loud speaker. The resultant vibrations in this diaphragm produce pressure vibrations in the air which are accurate counterparts of those which came from the original sound source. Thus listeners hear sound which is a virtual copy of the original. Since the photoelectric cell is actuated by light variations, sound is not reproduced when there is no vibration of the recording line of light upon the sound track and when, therefore, the latter is half exposed and half unexposed, as shown in that portion of Figure A marked b b' and c c'. But in the process of development into a negative of a film left in this condition through silence, scratches, abrasions, impurities, spots of silver and the like unavoidably appear — they are caused in the course of the developing process — in the otherwise transparent portions of the film, and when the developed film is then printed into a positive and used in connection with the photoelectric cell to reproduce sound, these imperfections cause a hissing or crackling noise called in the art "background noise." Its intensity is unfortunately greatest during what should be wholly silent periods.
A photoelectric cell consists of a glass bulb silvered internally. On this inner coating is deposited a thin layer of some metal such as potassium or selenium. In the center of the bulb there is arranged a collecting electrode supported on a wire sealed directly through the glass. A connection is made to the metal in the same way. The silver is removed from a certain area of the glass so as to allow light to enter the cell. Exposure of the metal to light generates an electric current. The function of the cell in respect of the art under discussion is thus to transform variations in light into variations in an electric current. 17 Encyclopaedia Britannica (14th ed. 1929, 1932) 792.
This production of background noise may be further explained as follows: ". . . These foreign particles, in passing the scanning aperture [of the photoelectric cell], interrupt the light beam in the same manner as a modulation wave, and so produce undesirable noise in contrast to the signal itself, which is producing the wanted sound. Since these particles are located at random throughout the clear area of the track, their power of modulation is proportional to the width of the clear track area. . . . Without modulation, that is, with no signal recorded on the track and without noise reduction, the widths of the clear and opaque areas are equal. The division line between these two areas is known as the zero line or base line, and is the axis of the recorded wave during modulation. . . . At low modulation levels the clear area width is excessive and as a consequence the foreign particles mentioned above are present in a relatively large extent, which results in a serious prominence of ground noise. The obvious remedy is the reduction of the excessive width of the clear portion of the track during low modulation levels. . . ." Noise Reduction, Fred Albin, Chapter IV, Motion Picture Sound Engineering (D. Van Nostrand Company, Inc., 1938).
Since the occasion of this background noise was the excessive width of the clear portion of the sound track in periods of silence, inventors sought to solve the problem of background noise by bringing the median or zero line of the sound track variations near to one side of the track during silent periods. We are thus brought to a discussion of the references Hewlett and Robinson, for these inventors in this manner accomplished the solution of this problem. Their methods were so similar that they can be discussed together. Hewlett and Robinson made two changes in the system of sound recording outlined above: First, by means of an initial mechanical adjustment of the oscillograph mirror c of Illustration I, they so fixed the transversely disposed line of light impinging upon the sound track that its recording end was at one edge of the track in periods of silence, rather than at the middle. Second, they adjusted the apparatus in such manner that the median or zero position of the recording end of the line of light was moved transversely across the track, that is to say, away from the edge, as sound was produced and increased in intensity. As a result the zero line became a curved line varying with sound amplitude, rather than a straight line down the middle of the sound track as in the prior art. The result of these adjustments is shown in Figure C of Illustration II, where the line a a' represents the position of the recording end of the line of light during periods of silence, and the curved line b × b' represents the transverse disposition of the zero line during periods when sound is being produced. The first adjustment was made in order that when no sound was being recorded the transparent area (represented by the narrow white strip between the lines a a' and c c' in Figure C) would be at a minimum, thus substantially eliminating the occasion of background noise. The second adjustment was made in order to accommodate to the sound track, so far as possible, as the intensity of sound increased, the peaks of the variations described by the recording end of the line of light.
There will be noted in the text the use of the phrases "zero line" of the sound track variations, "zero position" of the recording end of the line of light, "median line" of the sound track variations, and the like. This is the phraseology of the record. It is superficially confusing in that the word zero seems to be used in connection with both rest and motion and in connection with the position of the recording end of the line of light when at the edge of the track and when away from the edge and also when at the center. But the confusion will be dissipated if it be borne in mind that, as explained in footnote 7 above, by the zero line is meant the division line between the widths of clear and opaque areas upon the sound track [when developed into a negative] in periods of silence and the axis of the recorded wave during periods of sound. The term median line is usually used interchangeably with zero line to describe the axis of the recorded wave during periods of sound; the term zero position is applied to the position of the recording end of the line of light in periods of silence. The terms "average exposure" and "average position" are also used in the record as synonymous with "zero line" and "median line."
In order to accomplish the second adjustment, Robinson and Hewlett branched off through a rectifier, indicated in Illustration I, connected in parallel with the rest of the system, a part of the alternating current as it came from the microphone. This rectifier converted the alternating current into direct current, and the latter was returned to the system and passed on through the oscillograph. The direct current thus produced varied in volume in accordance with changes in amplitude of the sound and its effect upon the oscillograph was to cause the mean position of the same to move in a manner independent of the oscillograph oscillations caused by the sound modulated alternating current; and this independent motion produced by the direct current was such that as sound amplitude increased, the median position of the recording end of the line of light, which latter had been set by the first adjustment at the edge of the sound track in periods of silence, was carried transversely toward the opposite edge of the track. Upon this solution of the problem of eliminating background noise Robinson and Hewlett were granted their patents.
This was explained by the appellant Poulsen as follows: "Now, this rectified current, as it is rectified from the alternating current, will always be in accordance with the volume of the sound. If you have a weak sound it will be small, and if you have a more powerful sound it will be greater. So the end of the light line, or the recording edge of the light line will be removed, or I should say the average position of the recording edge of the light line will be removed towards the other side of the sound track. We may express that by saying that a variable bias was produced on the zero line."
But, unhappily, this solution of the problem of background noise produced another problem for the art. To understand that problem it is necessary to refer again to the art of sound recording and reproduction as it stood before Robinson and Hewlett. It will be seen from an examination of Figure A in Illustration II that some of the peaks of the variations of the recording end of the line of light fall outside the margin of the sound track. This was because it is necessary to keep the sound track somewhat narrow, since most of the film must be devoted to the visual images, and when especially loud sounds occur the wide variations produced by them will exceed the width of this narrow track. The effect of this was to cause a distortion of sound — referred to in the art as the introduction of harmonics. This is explained by the appellant Poulsen thus:
"In sound recording one is always trying to keep the maximum peak of the recorded sound inside the border of the sound track in order to avoid distortion, because if it is going beyond it will be masked off either in recording or in reproduction or both, so what is outside is not only not recorded, but on account of the sudden cut-off, the sharp cut-off, distortion will be introduced. Technically, it is expressed by saying that harmonics are introduced. Like in radio, if you overload a radio tube you cut off the maximum peaks and the sound is distorted.
"Now, whereas one will always try to avoid such cut-off, it is not in practice always possible to avoid it because the volume of sound may suddenly and unexpectedly go up. For instance, if an actor is shouting, or if you have a shot of a gun, or something like that."
In the art prior to Hewlett and Robinson, this distortion was not a serious obstacle to good sound reproduction because the zero line then ran down, and was never displaced from, the middle of the sound track; hence, when the variations produced by loud sounds exceeded the width of the track, the excess was equal on opposite sides, and the variations were thus cut off symmetrically. This is shown in Figure A of Illustration II. And where the cut-off is symmetrical the distortion of sound is not serious. But in the art as modified by Hewlett and Robinson, the cut-off of the extreme peaks of variations produced by loud sounds was asymmetrical — on one side only — as shown in Figure C. This was the necessary result of the two adjustments made by Robinson and Hewlett, the first of which, as above explained, set the recording end of the line of light at one edge of the track in periods of silence and the second of which carried the median line toward the opposite side as sound commenced and increased in amplitude. This distortion was harmful.
On this subject Poulsen testified: "In normal sound recording this did not produce a very harmful distortion because if the zero line was in the middle of the track the cut-off would be symmetrical, the sounds would be cut equally off on both sides of the sound track, and thereby some of the harmonics which are introduced are entirely eliminated."
Poulsen testified: "But in the noiseless recording invented by Robinson and Hewlett, as shown in Fig. D [sic], the cut-off will not be symmetrical. It will be asymmetrical. Only the peaks on one side will be cut off, and will be cut off double as much as they would be cut off in the normal recording, assuming that the over-shooting is the same in both cases.
"Now, this asymmetrical cut-off is causing a very harmful distortion which can be heard. So at the same time as this reduction of the background noise was obtained a rather harmful effect of distortion was introduced.
"Q. Could the average audience in a motion picture house discover or notice this distortion? A. Oh, yes.
"Q. Could you describe the nature of the distortion more specifically? A. As I said before, by the cutting off harmonics or overtones are introduced."
It was this problem of asymmetrical distortion that the appellants sought to solve. If they were to solve it usefully, they must not merely eliminate the distortion produced by asymmetrical cut-off of the light variations — they must also preserve the advance of Robinson and Hewlett in the elimination of background noise. They did both.
The appellants introduced into the art as it stood under Robinson and Hewlett the following changes: First, they mechanically adjusted the position of the oscillograph mirror so that the recording end of the line of light would fall at the middle of the sound track, at the point designated a in Figure D of Illustration II, in the absence of the flow of any current through the oscillograph loop. Second, by means of a battery f placed, as shown in Illustration I, between the rectifier and the point of re-entry of the rectified current into the system, they introduced into the system a direct current bias. The effect of this, operating alone, was to move the oscillograph mirror in such manner as to carry the recording end of the line of light close to, but not beyond, the edge of the sound track — to the point b in Figure D of Illustration II. Since this direct current entered the system from an external source it was not a sound modulated current, and therefore in this position of the recording end of the line of light the system was still free of any current actuated by the production of sound. Third, the appellants constituted such a relationship between the direct current bias and the rectified current (which, as above explained, operated, as a result of, and in accordance with, variations in sound amplitude, to move the oscillograph mirror independently of its oscillations) as to cause the variation in the rectified current, as sound was produced and increased in intensity, to oppose with constantly increasing effect the direct current bias; and this adjustment was so made that when loud sounds were produced the oscillograph mirror swung the zero line back to, but not beyond, the middle of the sound track. It swung back to the middle because the effect of the direct current bias to set it over to the edge, as above explained, was neutralized; it did not pass beyond the middle because the effect of the rectified current to displace the zero line as sounds increased in intensity was also at that point made no longer effective. In consequence, the median line of the variations described by the recording end of the line of light was kept within definitely predetermined limits: one, the edge of the track, the other, the middle. Thus the median line took the form b d b' in Figure D of Illustration II. The result of this was that where the amplitude of sound was such as to produce variations in the recording end of the line of light which were wider than the sound track, the excess was cut off equally on each side of the track, rather than unequally as in Robinson and Hewlett.
This was explained in more detail by Poulsen as follows: "We introduced in series with the rectifier a constant d.c. bias, represented in the figure by a battery, and we adjusted the connections in such a way that the variable bias produced by the rectifier would oppose the constant d.c. bias, so as to reduce the constant d.c. bias the greater the sound amplitudes are. At the same time we adjusted our oscillograph in a different way, whereas Robinson and Hewlett adjusted the oscillograph so that when it was current-less, when there was no current in it, the edge of the light line would be close to the edge of the track, and we adjusted the oscillograph so that when there was no current in it the end of the light line will be at the center of the track. That is one thing.
"Then our second step is to introduce this constant d.c. bias and by means of that to remove the light line so as to be near the one side of the sound track, and then when the sounds are coming in we reduced the constant d.c. bias by means of an opposing variable bias, so that the d.c. bias is gradually reduced, and we adjusted the values, so that just before any cut-off takes place the constant d.c. bias will be reduced to zero, whereby the oscillograph will again be moved towards the middle of the sound track, or I should say the light line will again be removed towards the middle of the sound track."
The question before the Patent Office was whether this solution by the appellants of the problem of asymmetrical distortion was patentable.
The claims of the appellants which define the method and apparatus just described are those numbered 72, 73, 74, 76, 77, 78 and 79:
"72. The method of producing a photographic sound record comprising the following steps: directing a light beam onto a moving film so that it receive a transversely disposed extremely narrow strip of light, variably exposing the film in accordance with the amplitude of each individual sound wave, controlling the light impinging on the film to vary the average exposure of the film as the volume of the sound to be recorded varies, and holding the variations of the average exposure between predetermined lower and upper limits.
"73. The method according to claim 72 and in which the lower limit of the average exposure variation is approximately zero.
"74. The method according to claim 72 and in which the upper limit of the average exposure variation is substantially midway between zero and full exposure.
"76. In a photographic sound recording system including a moving film, a source of light and a light flux controlling element, sound modulated alternating current means for controlling said element to expose the film in accordance with the amplitude of each individual sound wave, and biasing means for controlling said element to vary the average exposure of the film between a predetermined lower and a predetermined upper limit.
"77. The recording system according to claim 76 and in which the lower limit is approximately zero.
"78. The recording system according to claim 76 and in which the upper limit is substantially midway between zero and full exposure.
"79. The recording system according to claim 76 and in which a source of unidirectional current applies a constant bias on said element, and means for applying a rectified portion of the alternating current to reduce the bias substantially to zero as the sound volume increases from zero to a predetermined maximum." [Italics supplied]
The findings of fact of the trial court describe as follows that feature of the appellants' apparatus which the appellants claim constitutes invention:
"It is therefore a characteristic feature of plaintiffs' apparatus that the zero position of the oscillograph about which it oscillates is moved to said initial position[] when the sound reaches said predetermined maximum and maintains this position even if the sound volume increases above said maximum value. Consequently the cut-off of sound amplitudes exceeding this value will be symmetrical and harmful distortion will be avoided.
By this the trial court obviously means the center of the track.
"Plaintiffs' system will produce a negative sound record in which the average exposure varies between a lower limit which is substantially zero, and an upper limit which is substantially midway between zero and full exposure."
The findings also recognize that the patents to Robinson and Hewlett do not disclose "any means for limiting the average exposure of the sound record to any given proportion of its entire width." Notwithstanding this the trial court concluded, in respect of the appellants' claims above set forth, that no invention was involved.
The conclusions of law pertinent to the claims now under discussion read as follows:
"It was not invention to substitute means for limiting the average exposure of the sound record between a minimum of zero and a maximum midway between zero and full exposure for the mechanical and electrical biasing means disclosed in each of the patents, Hewlett No. 1,853,812 and Robinson 1,854,159."
"None of the plaintiffs' claims 49, 72, 73, 74, 76, 77, 78 and 79 distinguishes patentably from any one of the patents, Hewlett No. 1,853,812, Robinson No. 1,854,159 or Hanna No. 1,888,724."
"It was not invention to substitute means employing a direct current for biasing the oscillograph to a zero position near one edge of the sound track, as set forth in plaintiffs' claim 79, for the mechanical biasing means disclosed in each of the patents, Hewlett No. 1,853,812, and Robinson No. 1,854,159."
The findings of fact concern themselves solely with the apparatus claims. The conclusions of law, however, concern themselves with all of the claims, both apparatus and process, and the decree of the trial court had the effect of rejecting all of the claims which we are now discussing.
The initial rejection in the Patent Office of the claims now being discussed was explained in the "Examiner's Statement on Appeal before the Board of Appeals" thus:
"Claim 72 is rejected as fully met by the Hanna, Robinson and Hewlett patents. The claim reads directly on the references as it is composed of two steps, the recording of the sound record and the recording of the volume record. `Holding the variations of the average exposure between predetermined lower and upper limits' is a required normal adjustment of the recorder in order that the record may be confined to the record track area allotted therefor."
Claims 73 and 74 were rejected as depending upon claim 72. Claims 76, 77, 78, and 79, being the apparatus as distinguished from the process claims, were similarly disposed of.
The examiner did not in terms mention claim 78, but this was an apparent inadvertence.
The Board of Appeals followed the recommendation of the examiner against the claims now under discussion. But it did not adopt the examiner's idea that holding the variations of the average exposure between predetermined lower and upper limits in Poulsen and Petersen had been accomplished in Robinson and Hewlett by confinement of the record "to the record track area allotted therefor." The Board of Appeals said:
"Claim 72 has been rejected on the patents to Hanna, Robinson and Hewlett. Appellants concede that all the elements of this claim, except the last, are found in the prior art, but contend that the last element calls for the step of holding the variations of the average exposure, that is, the zero line between predetermined lower and upper limits because in all prior structures only one limit is set, the zero line being permitted to shift past the center line of the track. It is not seen that the fact that the zero line is permitted to shift past the center line indicates that it does not have a limit somewhere beyond the center line. The rejection of this claim appears to have been properly made."
The same view as that expressed by the examiner and the Board of Appeals is urged by the Commissioner on this appeal and was, we assume, also urged upon and accepted by the trial court. The Commissioner's brief states:
"Claim 72 recites that the variations of the average exposure are held between predetermined upper and lower limits. It does not state what these limits are and the Board of Appeals was of the view that in devices of the Robinson and Hewlett patents there must be limits somewhere to the average exposure, and that the claim does not distinguish patentably from either of the patents."
The rejection by both the trial court and the Patent Office was thus upon the ground of anticipation. The Patent Office seems to have raised no question in its own tribunals or in the trial court, and seems to raise none on this appeal, concerning utility.
We think that the appellants' improvement was not anticipated in Robinson and Hewlett and that the rejection by the trial court and the Patent Office was demonstrably the result of clear error in respect of the foundation for rejection. As appears from the foregoing quotations from the examiner's statement, from the decision of the Board of Appeals, and from the Commissioner's brief on this appeal, what decisively moved the Patent Office, and apparently the trial court, to rule that the appellants' process and apparatus are not patentable over Robinson and Hewlett was the concept that in Robinson and Hewlett there was some predetermined upper (as well as lower) limit beyond which the zero line of the variations described by the recording end of the line of light could not shift — whatever the amplitude of sound. But the description above of the construction and operation of the Robinson and Hewlett apparatus shows that this is not correct. In Robinson and Hewlett the oscillograph mirror is adjusted mechanically so that the recording end of the line of light is close to the margin of the track during periods of silence; and the rectified current reintroduced into the system and carried through the oscillograph loop operates to shift the median line of the variations back toward the center of the track as sound commences and increases in amplitude. See Figure C of Illustration II. But nothing is described in Robinson and Hewlett which will prevent the zero line from going as far past the center of the track as the amplitude of sound, however great, will carry it. Thus there was a predetermined lower limit, but not a predetermined upper limit, in Robinson and Hewlett. And since it was precisely the establishment of the upper limit for the median line at the middle of the sound track (keeping the lower limit nevertheless at the edge so as to avoid background noise) that was the heart of the appellants' solution of the problem of asymmetrical distortion, the error of the Patent Office and the trial court went to the core of the question of anticipation.
Apparently the Patent Office was misled, in respect of anticipation of the appellants' device and method by Robinson and Hewlett because, in its comparison of the appellants' device with that of Robinson and Hewlett, its attention was in the large concentrated upon the elimination of background noise rather than upon the solution of the problem of asymmetrical distortion. Of the latter the examiner's statement makes no mention. It states:
"The subject matter of this [the appellants'] application concerns the modification of a standard sound on film recorder, in order to eliminate or materially reduce the ground noise inherent in the film, which on reproduction of the record is objectionable in silent periods and when the intensity of the sound is low."
And in the discussion of the prior art the examiner's statement concerns itself largely with the problem of background noise elimination. While the decision of the Board of Appeals mentions the subject of distortion, it does so in connection with certain allowed claims whose method of avoidance of distortion was quite different from that of the claims with which we are immediately concerned here.
Claims 56 and 65 exemplify the distortion correction system with which the allowed claims are concerned.
"56. A system for recording sounds on a light-sensitive film comprising means for converting the sounds into a varying electrical current, a light source, a mirror oscillated by said current and reflecting a bundle of rays from said light source towards the film, means to concentrate said rays on the surface of the film to form a transversely disposed linear strip of light on same, a rectifier for rectifying a portion of said current, a thermionic amplifier, a wire loop carrying said mirror and inserted in the plate circuit of said amplifier and a resistance inserted in the circuit of the rectified current and connected across the grid circuit of the thermionic amplifier in order that the grid-bias should vary in accordance to the magnitude of the rectified current for the purpose of causing corresponding variations of the plate current of the amplifier, whereby the mean position about which the mirror is oscillated by the influence of the varying current into which the sounds are converted, is altered in dependency of the magnitude of the sound amplitudes so as to cause the zero-line of the recorded sound to be displaced transversely to the film from an initial position at one edge of the film towards the middle of the same at a rate increasing for increasing amplitudes of the recorded sounds, the conductivity of the last mentioned circuit being adjusted so as to cause that the variations of the rectified current, corresponding to increasing sound amplitudes, to be effected more quickly, and the variation of the same current corresponding to decreasing amplitudes to be effected more slowly than the variations in the sound amplitudes."
"65. A system for recording sounds on a light sensitive film comprising means for converting the sounds into a varying electrical current, a light source, means to concentrate a bundle of rays from said light source onto the surface of the film to form thereon a transversely disposed linear strip of light, means to cause said bundle of rays to vary so as to cause the exposure of the film to vary in accordance with the volume of the sound, a rectifier rectifying a portion of said varying electrical current, and a circuit through which the said rectified current influences the sound recording means to cause the said bundle of rays to oscillate dependent on the volume of the sound to be recorded, the electrical constants of at least one of the circuits through which the rectified current acts upon the sound recording member being of such a value that the rate of decrease of the current acting upon the recording means is greater during increase of the sound intensity than the rate of increase of the peak value of the alternating current." [Italics supplied]
The effect referred to in the italicized portion of these two claims is exemplified in Figure D of Illustration II at the points marked c and c. The allowed claims thus do not have to do with the subject matter of the rejected claims, that is, with the elimination of distortion by restricting the upper limit of the median line to the center of the sound track no matter what the amplitude of the sound.
Although the Commissioner's brief on appeal recognizes that "It is appellant's contention that inasmuch as the zero line of the oscillations cannot be displaced beyond the midline of the sound track the extremities of the oscillations will be equally cut off at both edges of the sound track and therefore no substantial distortion of the sound will be caused even if the oscillations are not confined entirely to the sound track," the brief immediately goes on to state that in both Robinson and Hewlett "the same result is sought as in the appellants' apparatus and method." And at the outset the brief of the Commissioner states that "The end sought by appellants and by other and prior inventors is to avoid the background noise by eliminating as far as possible the transparent areas." (Italics supplied) With the attention of the Patent Office thus upon the subject of the elimination of background noise, it was not unnatural that it should be misled in respect of the subject of anticipation. For it is true that insofar as the appellants in solving the problem of distortion, the asymmetrical cut-off, sought to save for the art the advance made by Robinson and Hewlett in eliminating background noise, they pursued the same method as Robinson and Hewlett, to wit, a moving of the recording end of the line of light close to the edge of the track in periods of silence so as to avoid a transparent area. Had the attention of the Patent Office been directed not merely to the elimination of background noise but to the whole of the appellants' accomplishment, i.e., the solution of the problem of asymmetrical distortion without sacrifice of the advance of the art in respect of background noise, it must have noted the crucial difference between the process and apparatus of the appellants and that of Robinson and Hewlett.
We think moreover that the improvement made by the appellants, simple though it may seem from the point of view of hind sight, was inventive in character. As said by Judge Learned Hand in Van Heusen Products v. Earl Wilson, D.C.S.D.N.Y. 1924, 300 F. 922, 929: "Is it not clear that the quality of a man's inventiveness must be tested by reconstituting the situation as it was in the light of the preceding history of the art?" When the situation which confronted the appellants at the time they commenced work upon the problem of distortion through asymmetrical cut-off is viewed in the light of the art as it then stood — under Robinson and Hewlett — the approach was apparently to an unsolvable problem. The appellants must not only venture upon an uncharted sea — no one had yet attempted solution of this problem of distortion — but also they must apparently sail between the Scylla of distortion and the Charybdis of background noise, for it was precisely the step found necessary by Robinson and Hewlett to the elimination of transparency — the occasion of background noise — that produced asymmetrical overlap and cut-off. Production of distortion thus seemed inescapably the price of the elimination of background noise. Putting this more concretely: If the appellants were to solve the problem of distortion and at the same time retain for the art the advance made by Robinson and Hewlett, they must go contrary to the teaching of the art as it stood under Robinson and Hewlett. For the latter, in order to eliminate transparency, set the oscillograph mirror so that initially, in the absence of sound, the recording end of the line of light was at the edge of the sound track; the appellants set the mirror so that initially, in the absence of sound, the recording end of the line of light was at the middle of the track. True, the appellants then swung the recording end of the line of light over to the edge, and swung the zero line back as sound was produced and increased in intensity. But their first step was nevertheless flatly contrary to that of Robinson and Hewlett. We think it took that boldness, clarity of vision and ingenuity characteristic of invention, as distinguished from mechanical skill, to attempt the voyage upon which the appellants set out and to sail it successfully so that at the end both the Scylla of distortion and the Charybdis of background noise were eliminated for the future of the art.
It is noteworthy that in three of the leading cases on the subject of invention this same boldness and success in solving an apparently unsolvable problem were involved. In Van Heusen Products v. Earl Wilson, referred to above, the problem was to preserve the comfort of a collar without starch and at the same time attain the shapeliness of a collar with starch. The invention provided a collar of multiple-ply interwoven fabric sufficiently stiff to maintain its shape without the employment of starch, but sufficiently pliable nevertheless to assume the necessary curvatures to fit the neck of a wearer without undue rigidity. The solution was through the use of the multiple-ply interwoven fabric. In Wach v. Coe, 1935, 64 App.D.C. 235, 77 F.2d 113, the problem was that of applying the power of a reciprocating steam engine and of a turbine — an exhaust steam engine receiving its steam from the reciprocating engine — to the same propellor shaft in a steamship. Since a reciprocating engine runs at low speed and with uneven torque and a turbine at high speed with even torque, the problem was apparently unsolvable to marine engineers. It was solved by introducing at the connection between the propellor shaft driven by the reciprocating engine and the shaft of the turbine three elements in series — a reduction gear, a hydraulic coupling, and a second reduction gear. In Smith v. Snow, 1935, 294 U.S. 1, 55 S.Ct. 279, 79 L.Ed. 721, the subject of invention was artificial incubation of eggs. Prior to Smith successful incubation had been thought dependent upon absence of artificial air currents, but without such currents of air incubation was restricted to small lots and to an unstaged process apparently required because of the endothermic and exothermic periods in the development of the embryo. Smith nevertheless introduced controlled air currents and staged incubation, and thus hatched eggs successfully in large lots. In each of these three cases the advance, simple as it was, was held patentable. Simplicity, indeed, is frequently a mark of invention. In re Reckford, Cust. Pat.App. 1933, 62 F.2d 842.
Moreover, it can hardly be said that the step taken by the appellants in eliminating distortion was less inventive in character than that taken by Robinson and Hewlett in eliminating background noise. While the rule of stare decisis does not apply to decisions of the Patent Office, there is a presumption of correctness in such decisions, and if what the appellants did is at least as inventive in quality as what Robinson and Hewlett did, the granting of a patent to the latter should be persuasive in respect of the right to a patent by the former. As the art stood when Robinson and Hewlett commenced to work upon the problem of background noise, it taught that background noise was occasioned by transparency and that transparency existed on half of the track during periods of silence since in periods of silence as well as of sound the zero line was at the middle of the track. It therefore might have been thought that the art taught Robinson and Hewlett that, in order to avoid background noise, the zero line must be moved to the edge of the track in periods of silence. But the Patent Office thought otherwise. If it was inventive in character, as the Patent Office held, for Robinson and Hewlett, in the then state of the art, to move the zero line to the edge of the track to avoid transparency and background noise in periods of silence, it was, we think, a fortiori, inventive for the appellants, after Robinson and Hewlett, to move the zero line back to but not beyond the center in order to avoid distortion during periods of sound.
Since in our view the findings and conclusions of the trial court and the rulings of the Patent Office were in clear error on the question of anticipation, and since we think it clear also that the appellants' advance was inventive in character, we are obliged to reverse in respect of claims 72, 73, 74, 76, 77, 78 and 79. Abbott v. Coe, 1939, 71 App.D.C. 195, 109 F.2d 449.
There remain for discussion claims 49 and 70. Claim 49 reads as follows:
"49. An apparatus for producing a photographic sound record of the variable width type comprising means for directing a beam of light onto a moving film so as to form a transversely disposed narrow strip of light on the film, means actuated by a sound responsive current for varying the length of the strip of light in accordance with the amplitude of the sound, a thermionic tube arranged so that its plate current actuates light varying means, means for applying a biasing potential to the grid of said tube, means actuated by a rectified portion of the sound responsive current for controlling said biasing potential in accordance with the volume of the sound so as to cause the plate current of the tube to vary in accordance with the volume of the sound, the apparatus being so adjusted that the zero line of the sound trace takes up such an initial position [] relatively to the sound track that, when the plate current attains a zero value, the zero line moves from the said initial position into the center line of the sound track, said apparatus including an oscillograph for directing the beam of light onto the film and for swinging the beam in accordance with the amplitude of the sound, in which the loop of the moving system of the oscillograph is connected in the plate circuit of the thermionic tube in series with two ohmic or inductive resistances, which are arranged one on each side of the loop, with a condenser connected across the said resistances and the loop, in order to prevent current of audible frequencies in the plate circuit from affecting the mean position of the moving system of the oscillograph." [Italics supplied]
The initial position here referred to is the edge of the track, to which the recording end of the line of light has been swung — from the center — by the direct current bias.
So far as the first italicized portion of claim 49 is concerned, the claim was rejected by the examiner and by the Board of Appeals upon the same ground as that upon which the claims first discussed were rejected, that is, upon the mistaken theory that in Robinson and Hewlett there was a predetermined upper limit to the transverse movement of the median line of the light variations. The findings of fact of the trial court do not mention claim 49 specifically. In conclusion of law No. 5, the claim is rejected with claims 72, 73, 74, 76, 77, 78 and 79, as not distinguishable patentably from Hewlett, Robinson or Hanna. This rejection is apparently upon the same ground as that which actuated the Patent Office. The views we have expressed above in respect of claims 72, 73, 74, 76, 77, 78 and 79 therefore require the conclusion that so far as the first italicized portion of claim 49 is concerned the rejection was improper.
As stated at the outset of this opinion, Hanna is not now relied upon by the Commissioner.
Claim 49 was rejected, so far as the second italicized portion is concerned by the examiner, the Board of Appeals and the trial court on the theory that it was not patentable over Hewlett, Robinson and Hanna. The rejection by the trial court, in conclusions of law Nos. 2 and 3, was apparently on the basis of the action of the Patent Office. The appellant Poulsen explained this portion of claim 49 in terms of Figure 4 of his drawings appearing in Illustration III below, in terms of Figure 1 of the Hewlett drawings shown in Illustration IV, and in terms of Figure 1 of the Robinson drawings appearing in Illustration V.
Poulsen's testimony respecting this part of claim 49 was:
"This is a rather restricted claim, referring to certain details of our circuit, which are not to be found in any of the cited patents. It refers to Fig. 4, where in the plate circuit of tube p two resistances ( q) are inserted in series with the oscillograph, and a condenser ( r) is connected in parallel to the two resistances and the oscillograph.
"The object of this arrangement is to eliminate from the oscillograph any trace of audible current which may have been left in the rectified current which is applied to the loop as a result of the rectification and the smoothing of that rectified current.
"When you rectify a current, it will have a wavy curve. It is consisting of all the half waves, for instance, all the positive half waves, of the alternating current, and therefore, it will be a pulsating current.
"Now, these variations would be audible if they were not smoothed out, and therefore a smoothing circuit is applied in connection with the rectifier, but there may be still some small traces of alternating current left, and it would be harmful to add them to the original alternating current in the oscillograph. If you did not eliminate them entirely, they would go into the record and produce some distortion, and therefore this arrangement has been added in order to entirely eliminate any trace of ripples in the rectified current.
"This is not found in any of the references cited in the case, and I believe there has been some misunderstanding in the Board of Appeals as to the object of this circuit.
"Q. The resistances and the inductances and the capacities with which the prior patentees deal, will you state how these are connected or in what portions of their circuits they are connected? A. They are all forming part of the smoothing circuit in connection with the rectifier.
"For instance, in the Robinson patent the inductances 40 in Fig. 1 are inserted in order to smooth out the ripples.
"And in Hewlett's patent, resistances 36-38 and the condenser 37 are inserted in order to smooth out the ripples.
"Also in Hanna an inductance 33 and condensers 34 are shown in the smoothing circuit, but they are all placed in the smoothing circuit.
"Q. In the smoothing circuit of the rectifier? A. In the smoothing circuit, yes, and they are all affecting the time constant of the rectifying circuit.
"Q. And will you state it again as distinguished from that, what your patent does? A. As distinguished from ours, which are inserted in the plate circuit in which the oscillograph is inserted, and are distinctly separate from the rectifying circuit." [Italics supplied]
We think it appears from this testimony that the only difference claimed for this filter or smoothing device is that it is inserted in the appellants' system separate from the rectifying circuit, whereas in Robinson and Hewlett it is a part of the same. We do not think it can be said that the Patent Office and the trial court were clearly mistaken in holding this not a patentable difference.
Claim 70 reads as follows:
"70. A system for recording sounds on a light-sensitive film comprising means for converting sounds into an alternating current, a light source, means to concentrate a beam of light from said light source onto the surface of the film so as to form thereon a transversely disposed linear strip of light, means to cause the exposure of the film to vary in accordance with the alternating current, a rectifier rectifying a portion of said alternating current, and a smoothing circuit through which the rectified current influences said sound recording means to cause the exposure to vary dependent on the volume of sound, the electrical constants of the rectifier and said smoothing circuit being so adjusted that the time constant of the total rectifying circuit is of smaller value when the rectifier is operative than when it is inoperative." [Italics supplied]
The italicized portion describes the claimed patentable advance. The claim was rejected by the trial court (in conclusion of law No. 6) and by both the examiner and the Board of Appeals upon the ground that it did not meet the demands of Rev.Stat. § 4888, 35 U.S.C.A. § 33. This requires the application for a patent to be "in such full, clear, concise, and exact terms as to enable any person skilled in the art or science to which it appertains, or with which it is most nearly connected, to make, construct, compound, and use the same. . . ." The examiner thought the claim vague since "a comparison between the time constants during operation and inoperation is meaningless because during periods of inoperativeness the time constant does not exist. Further, what constitutes the `Total rectifying circuit' is not specified." The Board of Appeals said that it was not clear to it what the claim meant. The appellants' brief attempts to justify the claim by saying:
"The italicized passage, which is concededly novel, means that the smoothing circuit is so adjusted that the time constant of the total rectifying circuit is of smaller value when the current is flowing through the rectifier than when no current is flowing through the rectifier."
Poulsen testified:
"That expression that the rectifier is operative or inoperative was held vague and meaningless by the Examiner. But it is well understood by those skilled in the art that a rectifier is operative when a current is flowing through it and inoperative when no current is flowing through it."
In respect of this aspect of claim 70 we are in accord with the views of the trial court and the Patent Office that it was not sufficiently definite to comply with Rev.Stat. § 4888. We therefore affirm the ruling below in respect of claim 70.
This case is therefore reversed in part, affirmed in part, and
Remanded to the trial court for further proceedings not inconsistent with this opinion.